Y/Pr/Pb modifying an old NTSC set?

[lnx64]

This is something I've been interested in. While it's fairly easy to RGB modify a TV with an on screen display (or closed caption decoder for that matter), older sets that don't have OSD's and analog tuners tend to seem, to me at least, like it could be possible to inject Component Y/Pr/Pb video internally somewhere, or at least S-Video.

I want to use my old Samsung set I used to use years and years ago again, but want at least S-Video or Component input.. The tuner will not be used again and bypassed completely since there's really no point to it anymore.

RGB would be nice, but I decided Component or S-Video because getting HDMI converters for that is far easier, and I can just use an HDMI switcher. Short of bypassing the tuner so I don't need a modulator, it would at least give it a FAR better picture than it ever had seen before by bypassing the comb filter entirely.

Samsung C9110MA. Uses a K20 chassis.

[freakaftr8]

I'm in the similar boat. Want to composite mod my Sony KV-2643R. RGB modify would be inputting the rgb signals into the neck board where the signals go from the mainboard I believe. I've heard of people attempting this with mixed results. pure RGB, but syncing would be done via mainboard where the IF video detector input goes.

Dont quote me on the IF part, but I believe its something like this.

[old_tv_nut]

Are you looking for comments or just posting a notice of your next project?

If you want comments, how about a link to a schematic?

[lnx64]

Originally Posted by old_tv_nut

Are you looking for comments or just posting a notice of your next project?

If you want comments, how about a link to a schematic?

Comments of course, I don't have the schematics, I was hoping someone might have them, which is why I listed the model of the TV and the chassis number. The purpose of this forum is for us to be able to help each other, not snip at them. That was kinda rude.

[lnx64]

Originally Posted by freakaftr8

I'm in the similar boat. Want to composite mod my Sony KV-2643R. RGB modify would be inputting the rgb signals into the neck board where the signals go from the mainboard I believe. I've heard of people attempting this with mixed results. pure RGB, but syncing would be done via mainboard where the IF video detector input goes.

Dont quote me on the IF part, but I believe its something like this.

Going to the neck board, wouldn't the voltages be different? I think RGB is 0.7v peak to peak. Even with that out of the way, I fear that route may end up with a picture slightly shifted.

[Electronic M]

S-video and RGB would probably be the easiest to implement. If you are inject S video just after the split the factory delay line will be right and RGB injection will also keep the 3 video signals in time with each other. Y/Pr/Pb I've seen someone try to add to a roundy...They didn't have the patience to make a custom Y delay and couldn't find an injection point that would give them the right delay so luminance lead chromanace in being scanned on the screen resulting in color fringing effects....

Some of those HDMI adapters had the ability to switch to RGB IIRC. You also might be able to find a HDMI to VGA converter that supports 480i which would give you sync outputs.

[lnx64]

Ahh, I forgot about the delay between luma and Pr and Pb for Component. Ok so S-Video honestly may be a better modification to this TV then, it's only a 19" going in a bedroom so I probably won't notice the difference anyway all that much, though for gaming that would be a nice bonus on my modified retro consoles and watching TV in a higher fidelity than originally intended for the set.

[old_tv_nut]

I'm not so sure there should be any delay between luma and Pr, Pb in any source, except for the effects of subsampled Pr, Pb sample siting. I would expect any delay due to Pr, Pb filtering to be corrected in the source.

Here's a paper on the possible variations. In any mpeg-2 source, I would expect to see the slight (one sample) luma delay shown (MPEG-2 4:2:0 sampling, far right at top of p.2), not a luma lead.

https://poynton.ca/PDFs/Chroma_subsampling_notation.pdf

[Electronic M]

Originally Posted by old_tv_nut

I'm not so sure there should be any delay between luma and Pr, Pb in any source, except for the effects of subsampled Pr, Pb sample siting. I would expect any delay due to Pr, Pb filtering to be corrected in the source.

Here's a paper on the possible variations. In any mpeg-2 source, I would expect to see the slight (one sample) luma delay shown (MPEG-2 4:2:0 sampling, far right at top of p.2), not a luma lead.

https://poynton.ca/PDFs/Chroma_subsampling_notation.pdf

Not in the source. I was talking about in propagation through injection points the TVs amplification stages.

[freakaftr8]

Originally Posted by lnx64

Going to the neck board, wouldn't the voltages be different? I think RGB is 0.7v peak to peak. Even with that out of the way, I fear that route may end up with a picture slightly shifted.

This is why you would sub in 75ohm resistors on the RGB IIRC

[old_tv_nut]

Originally Posted by Electronic M

Not in the source. I was talking about in propagation through injection points the TVs amplification stages.

Aha. In that case, it's a question of the demodulator and color difference stages bandwidth. Color bandpass amp will not be used, so the overall color difference delay will be less than with a composite or S-Video input. If you input the Y before the delay line, it will be delayed too much, and lag the chroma. If you input it after the delay line, it will lead the chroma.

Things that will affect chroma delay are deliberate filtering following the color demods (tube sets: shunt capacitors on the demod plates, series coils between the demods and color difference amplifiers, etc.). Explicit filter components can be removed (caps) or shorted (coils), since there is no need to filter out the subcarrier. However, this may still leave somewhat reduced chroma bandwidth (and some chroma delay) due to plate capacitance and large load values in the chroma stages.

If you can modify the color difference paths to have the same bandwidth as luma, the delay should be the same. Not guaranteed, but maybe possible.
Need to see the particular schematic in order to figure out whether it's worth a try.